本文目前僅提供英文版本。
Diligence on an industrial services company follows a well-worn path: quality of earnings, the WIP schedule, backlog, customer and general contractor concentration, safety and insurance, and management interviews. Each answers a real question. Almost none of it looks at the work itself, meaning the drawings the company bid from, the estimates it built, the changes it argued for and the files it handed over when the job was done.
Those documents sit on a file server and in estimators' folders, and rarely in the data room. They are worth asking for. A contractor's bid archive records how it actually estimates, job after job, in a form no management presentation can tidy up. This post covers what a sponsor can learn from it, how to get access without creating new problems, and how the findings become the first draft of the 100-day plan.
What the financials cannot show
The WIP schedule is the center of contractor diligence for good reason. Fade or gain between bid margin and final margin, job by job, shows whether the company prices work it can deliver. We covered why that matters for a buy-and-build thesis in The missing layer in the industrial roll-up.
What the WIP schedule cannot show is cause. A job that faded may have been underbid on quantities, lost money on a scope gap the estimate never covered, run badly in the field, or met a general contractor that refused legitimate change orders. Each cause has a different fix and a different meaning for the deal. Two targets with similar fade on paper can be very different businesses. One has an estimating problem that a shared take-off workflow can address in the first months of the hold. The other has a field execution problem that takes years.
Management interviews fill the gap with a narrative. It is usually sincere and often right. But it is one person's account of hundreds of jobs, and the documents hold the jobs themselves.
The WIP schedule says a job lost margin. The drawings, the estimate and the change log say where it went.
What the documents reveal
Read a sample of complete job files side by side and six patterns show up.
Estimating discipline. Does each estimate match the set it was priced from, including every addendum? Are quantities tied to sheets, or carried as totals with no trail? Do two estimators pricing similar buildings arrive at similar quantities, or does each have a private method? A disciplined shop leaves a trail from the drawing to the number. Others leave a number.
Scope habits. The scope letter, exclusions and qualifications show what the company habitually carries and what it pushes onto others. Controls, test and balance, insulation, firestopping, roof curbs, startup: each is a line someone owns. Compare the exclusions on won jobs with those on lost ones, and with the change orders that followed. A company that excludes the same items on every bid and still pays for them on many of its jobs has a habit worth knowing about.
Allowance patterns. Owner allowances are written into the bid documents, and every bidder carries them. Estimator allowances stand in for work nobody had time to count, and those are the ones to measure: how much of a typical estimate is plugs, factors and lumps rather than counts, in which categories, and by which estimators. We explain why in Bid on counts, not allowances.
Change-order recovery. Read the change-order log against the bid set and the construction set, and changes sort into two piles. Scope added after bid should be recovered, and how well it was shows the strength of the project managers' documentation. Scope that was on the bid documents all along cannot be recovered, and it shows what the estimate missed. Pending and rejected change orders on closed jobs show how the company handles the argument.
Close-out. Are the as-builts, operation and maintenance data and warranties in the job file, or does the file end at the last pay application? Close-out quality bears on retention collection, service work and repeat business with owners, and it is a cheap signal of how the company finishes a job.
How knowledge is distributed. Whose initials are on the estimates, and who reviewed them? Which bids went through the chief estimator, and how did the rest perform? The archive shows where the company's judgment actually sits, not only who says they hold it.
Questions to answer from documents, not interviews
Interviews are good at intent, culture and plans. Documents are better at distribution: how often, how much and by whom. A useful split:
| Question | Documents that answer it | What to look for |
|---|---|---|
| Is estimating repeatable? | Estimates set against their bid sets, across estimators | Quantities traceable to sheets, and similar methods across people |
| How much of the number is a guess? | Estimate workbooks and scope letters | The share of lines carried as allowances, factors or lumps, and where |
| Where does fade come from? | Bid set, construction set, change-order log, job-cost report | Overruns traced to missed scope, to quantities or to field labor |
| Does the company recover change? | Change-order log against the drawings and RFIs | Changes documented against the revision that caused them |
| Who is the estimating system? | Estimate authorship and review records | Concentration of bids, and of won margin, in one or two people |
| Does the thesis have a history? | Job files in the target segment | Completed work of the type the plan depends on, and how it closed |
The last row matters more than it looks. Many theses rest on a segment, such as hospitals, labs or data centers. The archive shows how many of those jobs the company has actually bid, won, built and closed, and how they performed.
Sampling the archive
Nobody reads every job in diligence. A good sample is small and deliberately uneven:
- Won and lost bids in the segment the plan depends on, so the estimates can be compared with the market's answer.
- High-fade and high-gain jobs from the WIP schedule, plus a few that closed on plan for contrast.
- Every active estimator, so the sample shows the company's method rather than one person's.
- Recent and older work, to see whether habits are improving or drifting.
For each job, ask for the complete file, not the estimate alone: the bid set and addenda as issued, the estimate workbook, the scope letter, the construction set, RFIs, the change-order log, the job-cost report and the close-out documents.
An illustrative example. Say a deal team requests twenty complete job files, with bid sets of a couple of hundred sheets each. Read by hand, that is weeks of a senior estimator's time, which a deal timeline rarely allows. Read into structured quantities with per-item confidence and checked by an engineer, the sample becomes comparable across jobs and estimators, and every finding traces to the sheet and region it came from.
What the sample cannot do is replace the people. It gives the interviews better questions. "Your lab estimates carry a large controls allowance. Why?" starts a different conversation from "Tell us about your estimating process."
Access and confidentiality
Drawings are not ordinary diligence material, and the seller's caution is often justified.
- The drawings may not be the target's to share. Contract documents usually belong to owners and their designers, and many contracts restrict how they are used. Some facilities, such as hospitals, data centers, utilities and process plants, carry security sensitivities of their own. Counsel on both sides should agree what can be shared, with whom and in what form.
- Pricing is competitively sensitive. When the sponsor already owns a platform that bids against the target, labor units, markups and bid results are exactly what competition law is concerned with. Expect a clean team, findings reported in aggregate, and no line-level pricing reaching the platform's estimators before close.
- The material needs a controlled place to live. A data room built for PDFs and spreadsheets is not built for reading drawing sets. Whatever environment is used should be isolated, encrypted in transit and at rest, governed by role-based access control and backed by an audit log of who saw and changed what. Operon runs each customer in an isolated tenant or on-premise, and we do not train AI models on customer content.
- Settle what happens if the deal does not close. Retention, return and deletion of the material should be agreed before the first file moves.
A seller usually grants a narrow, well-governed request faster than a broad one. Twenty complete job files under a clean-team protocol is an easier conversation than a request for the archive.
How findings feed the 100-day plan
Document diligence pays twice. It informs price and structure, and it writes the first draft of the 100-day plan.
On price and structure, findings move specific assumptions. If most of the estimating runs through a founder whose earnout ends in two years, the retention and transition terms should say so. We discuss that case in The retiring expert is a data problem. If the value creation plan counts on better estimating, the sample shows whether there is a method to improve or one to build.
On the plan, each finding becomes an action with an owner:
| Diligence finding | Action in the first 100 days |
|---|---|
| Heavy estimator allowances in one bid type | Make that bid type the first live take-off workflow |
| Recurring change orders for scope that was on the bid set | Turn those scope items into checks on every verified take-off |
| Estimating concentrated in one or two people | Have them verify the historical jobs that best show their method |
| Weak change documentation | Tie RFIs and change requests to the sheets and revisions behind them |
| Thin close-out files | Make as-builts and close-out records a condition of closing a job |
| A scattered or incomplete archive | Start the day-0 inventory from what diligence already mapped |
The sample also becomes the baseline. The allowance share, the causes of fade and the change recovery measured in diligence are what the add-on's progress is measured against after close. The phases themselves are in The first 100 days.
The work is in the files
A contractor's financial statements describe the results of its work. Its drawings, estimates, change orders and close-out files are the work. A well-chosen sample of them, read before signing, tells a sponsor how the company estimates, what it habitually misses, how it recovers change and who holds the knowledge, with evidence rather than anecdote. It is also the first step of integration, taken before the deal closes. Operon for Private Equity shows how the same layer carries from diligence through the hold.
